With the remarkable improvement in cure rates of childhood cancer over the last several decades, over 80% of children in the U.S. with cancer today become long-term (5+ years) survivors. This growing population of survivors, currently over 400,000 nationwide, reflects a highly vulnerable group of individuals with a high probability of experiencing adverse health-related and quality-of-life outcomes. To provide proper care for this population and inform the design of future treatment regimens, it is imperative to gain sound understanding of their long-term morbidity/mortality associated with specific therapeutic exposures, genetic profiles, sex and other demographic characteristics, and co-morbid medical conditions. Growing body of published literature exists addressing specific adverse outcomes and their associations with specific therapeutic exposures. While these association studies have provided key evidence utilized to develop follow-up care guidelines such as those of Children's Oncology Group, recognizing ?associations? (differences) is not sufficient for provision of individually-tailored follow-up care. Ability to ?predict? (prognose) is required to expand the impact of survivorship-based research and allow translation of observational research toward individualized-precision survivorship care. Capitalizing on existing strengths of the well-established and highly productive survivorship- based research program at St. Jude Children's Research Hospital, with currently available late effects outcomes data from direct clinical assessments and whole genome sequencing of germline DNA, the proposed project will apply state-of-the-art methods to construct and independently validate outcome-specific prediction models, based on a phenotype pathway/modeling framework for each outcome, incorporating genetic predictors with laboratory-based functional validation. We will undertake an aggressive program of survivorship focused research to address NCI's Provocative Question PPQ-7: How can prediction models be developed and used to identify patients at highest risk of treatment-related complications? By not restricting the proposed research to a small number of late effects outcomes or a specific diagnosis of childhood cancer, our aims are ambitiously set to extensively contribute to and meaningfully impact clinical practice. Our specific aims are to: (1) build individual-risk prediction models that have clinically-appropriate degrees of precision, for the following 11 late effects outcomes: meningioma; basal cell carcinoma; multiple subsequent neoplasms; cardiomyopathy; obstructive lung disease; restrictive lung disease; diabetes mellitus; oligo/azoospermia; primary hypogonadism; memory deficit; and executive function deficit; (2) validate them with independent validation cohorts; and (3) functionally/biologically validate the genetic elements in the risk prediction models. Upon completion of (1)-(3), we extend the model building, validation, and functional/biological validation work to 9 additional late effects: stroke; arrhythmia; growth hormone deficiency; hypothyroidism; central hypogonadism; processing speed deficits; attention deficits; hearing loss; and bone mineral density deficits.